Throughout my time at university, I have been passionate about exploring topics in international affairs. During my sophomore year, I enrolled in a International Journalism class, and my professor Shaheen - a bold woman of journalistic integrity mentored me to explore media landscapes through an expansive lens. As I now near the completion of my Masters’ program in Data Analytics and Computational Social Science, I am keen to go back to some of the foundations I learned about at college and look at them from a data and computational lens.
Last semester, as part of a group project, we qualitatively coded articles on Afghanistan and how the news was framed before, during and after the withdrawal of US troops. Narrowing my focus further, I am curious to explore how incidents of human rights are portrayed in media but also other non-profit publications that often are the basis for policy building and advocacy on a administrative scale.
This study conducts a topical analysis of coverage on human rights from around the globe. The original corpus itself includes raw text of over 14,000 human rights country reports published from four sources:Amnesty International (1974–2012), Human Rights Watch (1989–2014), the Lawyers Committee for Human Rights (1982–1996), and the United States Department of State (1977–2013). Given the computational limitations, I reduce the corpus to articles that were published after the year 2000 and run my analysis on 5728 observations. The overarching theme is to analyze the main topics that were talked about during this time frame and find targeted areas of research these kinds of models can complement. To do so, a structural topic model is incorporated into the analysis and based on goodness of fit, 13 topics and their contents are presented as primary results. While this analysis is exploratory in nature, it can serve as a launching point for various supervised and unsupervised learning methods to leverage computational techniques in the field of human rights and international relations.